The emergency steering override system failed because the operating pressure in its hydraulic control line was lowered by the partly open main steering by-pass distribution valve. Operating at the lowered pressure, the system failed to fully maintain the lock on the direction of propulsion units 3 and 4, which rotated slowly from the direction in which they had been set. All four units were initially set to turn ahead. However, as units 3 and 4 were at the offshore end of the ferry, they were able to draw more water than units 1 and 2. The flow of water to the latter units was restricted by the terminal. Therefore, the thrust from the less restricted offshore units (3 and 4) was greater than the thrust generated by units 1 and 2. When units 3 and 4 rotated from their setting, the after end of the ferry started to turn away from the terminal. It would appear that the safety rules regarding making the vessel fast by moorings when alongside were not observed. This may have been due to a lack of discipline on the part of the crew whose job it was to make the vessel fast, and to a lack of supervision by the vessel's officers. Although officers and crew are fully occupied at this time, the fundamental and simple standard task of making the vessel fast should not have been overlooked. It would also appear that it was ship's practice not to make the vessel fast when alongside and that this practice went undetected before this occurrence because the vessel's shoreside managers did not carry out safety audits to ensure that standard safety rules were observed. Because the moorings had not been made fast, the ferry, while slowly turning, departed the terminal, fortuitously at a time when no passenger or vehicle was embarking. If either the master or mate had been in the wheel-house at this time, the propulsion unit-induced movement of the ferry would have become apparent sooner and could have been corrected. The mate did not hear the master's call to come to the bridge to relieve him, probably because of ambient noise levels on the main loading deck. At the time, he was attending to his duties at the loading ramp, and had he responded to the master's call, he would not have been in a position to stop further passenger and vehicle traffic from embarking. The master had been on an altered work shift schedule since 22 April 1996, and his daily work period had increased from 8 to 10 hours. Unfortunately, the master lived on the mainland and the ferry overnighted on Wolfe Island. As a result, the master needed to join the ferry at 0100 on the mainland to travel to Wolfe Island to begin his 10-hour shift there at 0530. Thus, there was little remaining time at the end of each work shift for rest and sleep. At midnight, this work cycle would repeat itself. Analysis by the TSB found that this schedule, for persons working and travelling similar hours to the master, would not allow for adequate sleep. Over the four-day work period, a significant sleep deficit would accumulate. A sleep deficit can cause a reduced attention span and forgetfulness; a tendency to make riskier decisions; and a general decline in performance. The master's performance in this instance is consistent with the effects of fatigue in that he did not ensure that the bridge was crewed during his absence and he did not return to the bridge urgently. The vessel was not made fast to the terminal and was left unattended with the engines running, and with the wheel-house and engine controls accessible to the public.Analysis The emergency steering override system failed because the operating pressure in its hydraulic control line was lowered by the partly open main steering by-pass distribution valve. Operating at the lowered pressure, the system failed to fully maintain the lock on the direction of propulsion units 3 and 4, which rotated slowly from the direction in which they had been set. All four units were initially set to turn ahead. However, as units 3 and 4 were at the offshore end of the ferry, they were able to draw more water than units 1 and 2. The flow of water to the latter units was restricted by the terminal. Therefore, the thrust from the less restricted offshore units (3 and 4) was greater than the thrust generated by units 1 and 2. When units 3 and 4 rotated from their setting, the after end of the ferry started to turn away from the terminal. It would appear that the safety rules regarding making the vessel fast by moorings when alongside were not observed. This may have been due to a lack of discipline on the part of the crew whose job it was to make the vessel fast, and to a lack of supervision by the vessel's officers. Although officers and crew are fully occupied at this time, the fundamental and simple standard task of making the vessel fast should not have been overlooked. It would also appear that it was ship's practice not to make the vessel fast when alongside and that this practice went undetected before this occurrence because the vessel's shoreside managers did not carry out safety audits to ensure that standard safety rules were observed. Because the moorings had not been made fast, the ferry, while slowly turning, departed the terminal, fortuitously at a time when no passenger or vehicle was embarking. If either the master or mate had been in the wheel-house at this time, the propulsion unit-induced movement of the ferry would have become apparent sooner and could have been corrected. The mate did not hear the master's call to come to the bridge to relieve him, probably because of ambient noise levels on the main loading deck. At the time, he was attending to his duties at the loading ramp, and had he responded to the master's call, he would not have been in a position to stop further passenger and vehicle traffic from embarking. The master had been on an altered work shift schedule since 22 April 1996, and his daily work period had increased from 8 to 10 hours. Unfortunately, the master lived on the mainland and the ferry overnighted on Wolfe Island. As a result, the master needed to join the ferry at 0100 on the mainland to travel to Wolfe Island to begin his 10-hour shift there at 0530. Thus, there was little remaining time at the end of each work shift for rest and sleep. At midnight, this work cycle would repeat itself. Analysis by the TSB found that this schedule, for persons working and travelling similar hours to the master, would not allow for adequate sleep. Over the four-day work period, a significant sleep deficit would accumulate. A sleep deficit can cause a reduced attention span and forgetfulness; a tendency to make riskier decisions; and a general decline in performance. The master's performance in this instance is consistent with the effects of fatigue in that he did not ensure that the bridge was crewed during his absence and he did not return to the bridge urgently. The vessel was not made fast to the terminal and was left unattended with the engines running, and with the wheel-house and engine controls accessible to the public. The day before the occurrence, the engineer on duty did not completely close the distribution valve controlling the hydraulic pressure for the emergency steering override for propulsion/steering units3and4. The effect of the incompletely closed valve was that units 3 and4 could not be reliably locked in the emergency steering override position. While the WOLFEISLANDERIII was embarking passengers and vehicles at the Marysville terminal, the master locked units 3 and 4 in override and left the wheel-house unattended. The ferry was not secured to the terminal by moorings. During the master's absence from the wheel-house, propulsion/steering units 3 and4 rotated from the locked position, causing the ferry to move slowly off the terminal. The mate acted quickly to stop passengers and vehicles from embarking. By the time he also regained control of the ferry, she was about 45 m off the terminal. On 22 April 1996, the MTO had revised the work schedule for the shift crews, and three positions had been eliminated, increasing the normal number of hours worked daily from 8 to 10. During the four-day work period of the new schedule, a crew member living on the mainland is required to catch the 0100 ferry to Marysville to be ready to begin his 10-hour shift at 0530, leaving little time for rest and sleep. It is likely that the master was suffering from fatigue which may have adversely affected his judgement and performance.Findings The day before the occurrence, the engineer on duty did not completely close the distribution valve controlling the hydraulic pressure for the emergency steering override for propulsion/steering units3and4. The effect of the incompletely closed valve was that units 3 and4 could not be reliably locked in the emergency steering override position. While the WOLFEISLANDERIII was embarking passengers and vehicles at the Marysville terminal, the master locked units 3 and 4 in override and left the wheel-house unattended. The ferry was not secured to the terminal by moorings. During the master's absence from the wheel-house, propulsion/steering units 3 and4 rotated from the locked position, causing the ferry to move slowly off the terminal. The mate acted quickly to stop passengers and vehicles from embarking. By the time he also regained control of the ferry, she was about 45 m off the terminal. On 22 April 1996, the MTO had revised the work schedule for the shift crews, and three positions had been eliminated, increasing the normal number of hours worked daily from 8 to 10. During the four-day work period of the new schedule, a crew member living on the mainland is required to catch the 0100 ferry to Marysville to be ready to begin his 10-hour shift at 0530, leaving little time for rest and sleep. It is likely that the master was suffering from fatigue which may have adversely affected his judgement and performance. The WOLFEISLANDERIII accidentally departed from the terminal because the emergency steering override system failed due to an incompletely closed valve in the engine-room. Propulsion/steering units 3 and 4 turned slowly from the direction in which they had been set, causing the ferry to swing off the terminal. Contributing to the occurrence was the fact that the ferry was not secured to the terminal, the wheel-house was not crewed, and the master's judgement may have been affected by an accumulated sleep deficit caused by recent changes in his work schedule.Causes and Contributing Factors The WOLFEISLANDERIII accidentally departed from the terminal because the emergency steering override system failed due to an incompletely closed valve in the engine-room. Propulsion/steering units 3 and 4 turned slowly from the direction in which they had been set, causing the ferry to swing off the terminal. Contributing to the occurrence was the fact that the ferry was not secured to the terminal, the wheel-house was not crewed, and the master's judgement may have been affected by an accumulated sleep deficit caused by recent changes in his work schedule.